How Did Data from 3 Million Journeys Become Decisions?

A rail operator had data. Plenty of it. The real problem was how long it took for that data to reach the moment of decision.

In a large-scale rail operation handling about 3 million journeys a day, journey data, field inspections, performance indicators and operational records were kept in different systems.

Enterprise applications, databases, analytics infrastructure, reporting tools and mobile field apps each did their own job. But the answers to the questions that mattered most to operations were not in one place:

• Which lines are losing performance?
• Which stations are having problems?
• How does field data affect daily operations?
• Which indicators call for action?

The data existed. But bringing together and making sense of data spread across different systems took time. So the problem wasn’t creating a new data source; it was making existing data ready for decisions.

Building a shared data layer without replacing systems

The solution was not to tear out existing systems and rebuild them. Enterprise applications, the SAP infrastructure and its dozens of modules, various databases, reporting tools and field applications all kept running.

Butterfly sat on top of these systems and brought data from different sources together in a shared structure. Data held in different systems came together in a single decision layer.

The core idea was this: instead of replacing existing technology investments, connect the value that existing systems already produce. As a result, the organization could reach data held in different systems through a single operational view.

The 27-page report is now ready every morning

For operations teams, the daily report isn’t just a document. It’s the key source of information that sets how the day starts. In this project, the 27-page daily operations report began to be generated automatically every night.

Instead of spending the morning collecting data from different systems, merging files and preparing the report, the team can focus directly on the results. The process went from:

• Data collection → report preparation → review
• Data → ready report → evaluation → action

From 30 minutes to 3 seconds

Bringing the data together in one structure changed not only reporting but also query speed. Queries that used to take about 30 minutes now return in about 3 seconds.

The difference goes beyond a technical performance metric. If getting an answer takes half an hour, that question is often put off in the flow of operations. If it can be answered in a few seconds, the same question becomes a natural part of daily decision-making. As the time to reach data gets shorter, so does the distance to a decision.

100,000 rows of field data become performance indicators the same day

Rail operations produce field data constantly. Station inspections feed about 100,000 rows of data into the system every day. But simply recording raw data isn’t enough. The real value comes when that data turns into indicators that mean something for operations.

Field data can now feed into these indicators on the same day:

• Line performance
• Station performance
• Staff performance

This shortens the time between something happening in the field and the result showing up on the management screen. The field → data → analysis → decision chain speeds up.

6,000 records synced in 5 seconds

One of the biggest obstacles to different systems working together is syncing data at the right time and in the right form. In this setup, 6,000 records can be synced in about 5 seconds. Information created in different sources can be updated within the shared view that operations need.

The approach matters here: the goal isn’t to turn all systems into one system. The goal is to let different systems talk about the same operation.

Instead of waiting on IT for a report, users build their own screens

Reporting is another area where data access has been democratized. Operations teams can build the dashboards they need without writing code. They can also ask questions of the data in natural language to get the information they need.

So there’s no need to open a new report request or wait for the technical team for every new question. IT’s role shifts from producing reports to providing the infrastructure that lets the organization use data more effectively. Operations teams use the information they need directly.

The real transformation is in decision speed, not data volume

Processing data from millions of journeys certainly matters in this example. But the real issue isn’t the size of the data. Large organizations already have data:

• It’s in SAP.
• It’s in databases.
• It’s in field applications.
• It’s in reporting systems.

The real question is: how close does this data get to the moment a decision has to be made?

If an operations report is ready in the morning, the team can start the day making decisions instead of collecting data. If a 30-minute query returns in 3 seconds, questions that were put off can be asked again. If 100,000 rows of field data become performance indicators on the same day, the distance between operations and management can shrink. And if different systems can talk to each other, a new decision layer can be built while keeping the technology investment the organization has made over the years.

Because digital transformation doesn’t always mean replacing systems. Sometimes the real transformation is making the data your existing systems already produce faster, more accessible and more meaningful.

Outcome

In an operation handling 3 million journeys a day, data flows constantly. What creates value isn’t the flow of data, but turning it into the right decision at the right time.

Butterfly brings different data sources together in a shared structure without replacing existing enterprise systems, helping organizations get results from operational data faster.

Not more data. Faster decisions.